1 00:00:01,880 --> 00:00:04,480 Hey everybody. Welcome back to the Elon Musk 2 00:00:04,480 --> 00:00:07,040 Podcast. This is a show where we discuss 3 00:00:07,280 --> 00:00:11,480 the critical crossroads, the Shape, SpaceX, Tesla X, The 4 00:00:11,480 --> 00:00:17,080 Boring Company, and Neurolink. I'm your host Will Walden, Elon 5 00:00:17,080 --> 00:00:21,400 Musk's ex AI is developing a internal project called Macro 6 00:00:21,400 --> 00:00:26,640 Hard that aims to recreate Microsoft's core system 7 00:00:26,680 --> 00:00:32,159 ecosystem using AI alone. It goes beyond a typical 8 00:00:32,159 --> 00:00:36,320 research experiment and instead reflects Musk's intent to 9 00:00:36,320 --> 00:00:41,360 simulate, replicate, and eventually compete with 10 00:00:41,360 --> 00:01:03,240 Microsoft head to If you have a podcast player 11 00:01:03,280 --> 00:01:06,880 that has comments, I want to let you know that you're allowed to 12 00:01:07,600 --> 00:01:10,080 voice your opinion. I want you to voice your opinion 13 00:01:10,080 --> 00:01:14,080 on this. Is it a smart idea to allow AI 14 00:01:14,080 --> 00:01:19,040 to replicate code and also create more code in fix bugs on 15 00:01:19,040 --> 00:01:22,720 the fly? For enterprise businesses, that 16 00:01:22,720 --> 00:01:26,240 means big business, big, big business. 17 00:01:26,400 --> 00:01:29,640 I'm not talking 10 people, 20 people, 30 people. 18 00:01:29,960 --> 00:01:32,000 We're talking 10s of thousands of people. 19 00:01:32,240 --> 00:01:35,360 Would you trust your 10s of thousands of people in your 20 00:01:35,360 --> 00:01:40,920 organization to use AI software? Or do you want the homemade 21 00:01:40,920 --> 00:01:45,960 version where people create the software and then if you have a 22 00:01:45,960 --> 00:01:49,880 bug, you tell it to somebody and then that somebody tells it to 23 00:01:49,880 --> 00:01:53,960 somebody else and it goes down the chain instead of just robot, 24 00:01:54,040 --> 00:01:56,960 you know, and AI writing all the code. 25 00:01:57,920 --> 00:02:04,240 It's very interesting. So Elon Musk instructed XAI 26 00:02:04,240 --> 00:02:08,479 employees to use AI models to rebuild Microsoft Office, 27 00:02:09,240 --> 00:02:16,200 Microsoft Excel, Word, Teams, Windows, and even GitHub and 28 00:02:16,240 --> 00:02:20,520 PowerShell. Musk selected the name Macro 29 00:02:20,520 --> 00:02:23,960 Hard as a deliberate play on Microsoft's branding. 30 00:02:24,240 --> 00:02:28,360 Microsoft He expects the project to eventually challenge 31 00:02:28,360 --> 00:02:31,480 Microsoft's stronghold over software that underpins the 32 00:02:31,480 --> 00:02:34,080 majority of modern enterprises and governments. 33 00:02:34,320 --> 00:02:38,120 Do you want an AI in charge of your software that you use on a 34 00:02:38,120 --> 00:02:40,240 daily basis? They can see everything that you 35 00:02:40,240 --> 00:02:43,200 do. They know exactly what you do, 36 00:02:43,240 --> 00:02:49,080 all your keystrokes. Could they be tapping into all 37 00:02:49,080 --> 00:02:52,160 of your movements? Yeah, OK, that's just a 38 00:02:52,160 --> 00:02:55,120 conspiracy theory that was just to be kind of, you know, weird 39 00:02:55,120 --> 00:02:56,920 about it. But think about it. 40 00:02:57,120 --> 00:03:01,680 If you're tapped in in the cloud all the time anyway, could they 41 00:03:01,680 --> 00:03:05,360 track everything you do to, quote, make their software 42 00:03:05,360 --> 00:03:08,040 better? Could it be the end user 43 00:03:08,040 --> 00:03:12,640 license, you know, agreement that you have to sign off on 44 00:03:12,640 --> 00:03:16,160 that everything you do has to be tracked by this AI. 45 00:03:16,560 --> 00:03:20,240 And do you want XAI in particular to be in charge of 46 00:03:20,240 --> 00:03:26,240 that software? XAI is going to be used, Grok 47 00:03:26,240 --> 00:03:30,160 and other XAI models to execute real world software workflows 48 00:03:30,160 --> 00:03:32,760 without human engineering. Basically going to be able to 49 00:03:32,760 --> 00:03:37,920 prompt this thing and say I want to, I want these specific 50 00:03:37,920 --> 00:03:40,240 things, these specific features, I do it all the time. 51 00:03:40,240 --> 00:03:43,960 Specifications you, you write a spec sheet and then you have 52 00:03:44,080 --> 00:03:50,840 something like Grok or any other AI program you know that that 53 00:03:50,840 --> 00:03:54,560 they want to use. They can build the software. 54 00:03:54,560 --> 00:03:58,280 I can build software right now. Things that would take me that 55 00:03:58,280 --> 00:04:01,800 would have taken me weeks to build, even a month to build. 56 00:04:02,720 --> 00:04:11,040 I can instruct my AI companion to build something for me at 57 00:04:11,040 --> 00:04:13,160 lovable. Lovable dot dev. 58 00:04:14,840 --> 00:04:18,760 I did something the other day. Here's an example of what I did, 59 00:04:19,519 --> 00:04:21,560 and this is a very simple example. 60 00:04:21,560 --> 00:04:24,120 I had a spec sheet. I knew exactly what I wanted, 61 00:04:24,960 --> 00:04:28,440 knew exactly how it was going to run from the first user 62 00:04:28,440 --> 00:04:31,120 interaction on the homepage. It was a web app first 63 00:04:31,120 --> 00:04:35,600 interaction on the homepage all the way through the process of 64 00:04:35,600 --> 00:04:40,280 check out etcetera. And when I told the AI to write 65 00:04:40,280 --> 00:04:45,080 it, he told me this would be the phases, phase one, phase two, 66 00:04:45,120 --> 00:04:47,080 you know, five phase. I think it was 5 or 6 phases, 67 00:04:47,080 --> 00:04:51,480 I'm not sure. But eventually about two hours 68 00:04:51,480 --> 00:04:56,960 later, we had a fully functional MVP of the software that would 69 00:04:56,960 --> 00:04:59,400 have taken me two weeks to write probably. 70 00:04:59,880 --> 00:05:01,480 And then of course, that's the MVP. 71 00:05:02,360 --> 00:05:04,440 And that actually works through Stripe. 72 00:05:04,680 --> 00:05:06,960 There's e-commerce functionality, e-commerce pages. 73 00:05:07,280 --> 00:05:08,960 All the templates were there, everything was there. 74 00:05:08,960 --> 00:05:13,920 Everything was built out without using any of my hands to code 75 00:05:13,920 --> 00:05:15,560 it. I didn't use anything. 76 00:05:16,000 --> 00:05:21,960 I used Tailwind, Veet, React, etcetera to build this thing out 77 00:05:22,680 --> 00:05:28,720 and it was pretty damn good. So building something like Teams 78 00:05:29,960 --> 00:05:33,960 for an AI model, not going to be that hard. 79 00:05:33,960 --> 00:05:39,280 I, I'm assuming they're going to have this done by the end of 80 00:05:39,280 --> 00:05:41,720 this year, at least a really good model of it. 81 00:05:42,320 --> 00:05:46,600 Because if you can use AI, because also Microsoft already 82 00:05:46,600 --> 00:05:50,120 built this that, you know, use, building the code to make 83 00:05:50,120 --> 00:05:54,280 Windows like a Windows 11 clone is going to be a little bit more 84 00:05:54,280 --> 00:05:57,840 difficult. But something like an app like 85 00:05:58,200 --> 00:06:02,800 Microsoft Word, that's not going to be that difficult compared to 86 00:06:03,440 --> 00:06:07,200 say, a full operating system. Something like Teams, not as 87 00:06:07,200 --> 00:06:08,760 difficult as a full operating system. 88 00:06:09,000 --> 00:06:13,000 They want to use make PowerShell, Excel, Microsoft, 89 00:06:13,280 --> 00:06:17,520 basically Microsoft Office. They're not going to copy it 90 00:06:17,520 --> 00:06:19,760 exactly though. They're going to be able. 91 00:06:19,760 --> 00:06:22,240 They're just going to do this to demonstrate that EI systems can 92 00:06:22,240 --> 00:06:25,640 replicate the same capabilities through reasoning and code 93 00:06:25,640 --> 00:06:30,080 synthesis. So the lead devs will tell it 94 00:06:30,080 --> 00:06:32,400 what to do. We need something similar to 95 00:06:32,400 --> 00:06:35,400 Microsoft Word, but maybe they're going to do something 96 00:06:35,400 --> 00:06:40,440 like I did with Lovable and they were like, just ask it a 97 00:06:40,440 --> 00:06:42,680 question. How do you think the software 98 00:06:42,680 --> 00:06:45,760 would be better? I did that throughout the 99 00:06:45,760 --> 00:06:48,920 development cycle in at points. How would you make this software 100 00:06:48,920 --> 00:06:51,960 better? And tell me the steps you would 101 00:06:51,960 --> 00:06:56,640 do, why it's better, how it's better, and then give me 102 00:06:56,640 --> 00:07:01,560 examples of all of those things. And it did in most of the time, 103 00:07:01,640 --> 00:07:05,200 probably 99% of the time. It was really excellent at 104 00:07:05,360 --> 00:07:10,000 making those decisions. The user flow was at the 105 00:07:10,000 --> 00:07:12,840 beginning, it was an MVP. So it was very basic. 106 00:07:13,240 --> 00:07:17,720 User flow was just like point A, point B, Point C, you know, And 107 00:07:18,920 --> 00:07:25,360 they had other ways to flow the user to the end cycle, to the 108 00:07:25,360 --> 00:07:30,760 buy cycle in just fast and furious ways to get the job 109 00:07:30,760 --> 00:07:32,040 done. It was crazy. 110 00:07:32,040 --> 00:07:35,280 And that's just lovable. Could you imagine something like 111 00:07:35,280 --> 00:07:42,000 Grok who has all of the data? Or you know, I've used other 112 00:07:42,800 --> 00:07:45,080 apps too. I'm a front end developer, 113 00:07:45,200 --> 00:07:52,080 Claude, great at coding ChatGPT, it's pretty good, but it's, I 114 00:07:52,080 --> 00:07:55,080 think Claude is a little bit, I don't know, for the stuff that I 115 00:07:55,080 --> 00:08:01,320 do, it's a little bit better, but have Xai behind it, have 116 00:08:01,640 --> 00:08:04,680 Elon and Grok behind it. It's going to be something, you 117 00:08:04,680 --> 00:08:08,360 know, something ridiculous, I'm sure. 118 00:08:10,440 --> 00:08:16,480 The other thing is this is an internal target includes 119 00:08:16,480 --> 00:08:18,600 recreating Excel with AI agents alone. 120 00:08:19,320 --> 00:08:22,560 And those AI agents understand spreadsheet logic. 121 00:08:22,640 --> 00:08:26,240 They'll be able to generate formulas, suggest corrections, 122 00:08:26,320 --> 00:08:29,040 all without using any existing code. 123 00:08:29,040 --> 00:08:33,039 They're going to create the code completely from scratch, not 124 00:08:33,039 --> 00:08:37,679 going to use any libraries. Another team is focused on 125 00:08:37,679 --> 00:08:42,360 building a fully AI generated OS layer with basic window 126 00:08:42,360 --> 00:08:45,880 management, file storage, and application support. 127 00:08:46,280 --> 00:08:49,760 Again, completely from scratch, no frameworks. 128 00:08:50,200 --> 00:08:54,200 And if this is successful, Elon would be the first to show the 129 00:08:54,200 --> 00:08:58,160 general purpose AI can autonomously generate complex 130 00:08:58,160 --> 00:09:00,840 commercial software from first principles. 131 00:09:02,840 --> 00:09:04,880 No part when they start, there's nothing. 132 00:09:05,120 --> 00:09:09,560 It's just, you know, C++, you know C#, you know whatever 133 00:09:09,560 --> 00:09:14,360 language it is and you build this thing because you we're 134 00:09:14,360 --> 00:09:16,280 going to give you the prompt to build it. 135 00:09:16,400 --> 00:09:21,880 Crazy macro hard is not being built for external users right 136 00:09:21,880 --> 00:09:23,640 now. It will eventually, I'm sure. 137 00:09:24,440 --> 00:09:27,720 I'm sure they'll license this out or find a way to make money 138 00:09:27,720 --> 00:09:30,000 from it. The internal documentation 139 00:09:30,000 --> 00:09:34,840 frames the project is a stress test for grok in the bet. 140 00:09:34,840 --> 00:09:38,440 The general EI can reproduce software systems through self 141 00:09:38,440 --> 00:09:40,880 learning. That's how we get Skynet, man. 142 00:09:41,800 --> 00:09:44,160 This is the software builds the software. 143 00:09:45,120 --> 00:09:46,800 The part builds the part That's crazy. 144 00:09:47,440 --> 00:09:51,040 The initiative emerged shortly after Open A is growing presence 145 00:09:51,080 --> 00:09:53,880 in enterprise tools became clear, including its 146 00:09:53,880 --> 00:09:58,120 partnerships with Microsoft. Elon is, I wouldn't say jealous 147 00:09:58,120 --> 00:10:00,880 of Sam Altman. He's competing. 148 00:10:01,000 --> 00:10:07,480 He's competing directly with Sam and ChatGPT and Open AI. 149 00:10:07,680 --> 00:10:12,920 I think Elon is a little behind the game at this point. 150 00:10:13,840 --> 00:10:16,400 Grok chatbot, pretty cool with some stuff. 151 00:10:17,800 --> 00:10:23,200 It's all right, but it goes off the rails sometimes and open AI 152 00:10:23,200 --> 00:10:24,400 as well. It goes off the rails a little 153 00:10:24,400 --> 00:10:26,000 bit. Open AI. 154 00:10:26,120 --> 00:10:32,280 ChatGPT 5 wasn't such a big thing as everybody thought it 155 00:10:32,280 --> 00:10:35,960 was going to be iterative. You know, we had between 3:00 156 00:10:35,960 --> 00:10:39,720 and 4:00, ChatGPT, it was a huge upgrade. 157 00:10:40,480 --> 00:10:43,800 And everybody thought between 4:00 and 5:00 was going to be 158 00:10:43,800 --> 00:10:48,880 this massive gigantic switch, but it wasn't. 159 00:10:49,160 --> 00:10:51,200 Yeah. It was like, they're just fine 160 00:10:51,200 --> 00:10:57,800 tuning, ironing out some bugs, fixing some stuff up, making it 161 00:10:57,800 --> 00:11:02,880 a more solid platform before they move forward to GPD 6 and, 162 00:11:03,040 --> 00:11:05,520 you know, all the AIS after that. 163 00:11:07,720 --> 00:11:10,480 So Microsoft Open AI, they're working together, of course. 164 00:11:10,480 --> 00:11:16,880 And Elon thinks, well, I can do it better because he's also 165 00:11:16,880 --> 00:11:20,280 already criticized Microsoft's deep ties with Open AI. 166 00:11:20,280 --> 00:11:26,800 He doesn't believe that an quote open source AI company like Open 167 00:11:26,800 --> 00:11:31,960 AI should have funding from a closed source company like 168 00:11:31,960 --> 00:11:34,920 Microsoft. And Microsoft's invest billions 169 00:11:34,960 --> 00:11:39,000 in Open AI and tightly integrated its models into 170 00:11:39,000 --> 00:11:43,040 Office. And Musk has expressed concerns 171 00:11:43,320 --> 00:11:47,080 in the past about the influence Microsoft holds over Open AI's 172 00:11:47,080 --> 00:11:50,480 operations. Internally, XAI researchers view 173 00:11:50,480 --> 00:11:54,560 macro hard as a response to what they consider model capture, 174 00:11:54,800 --> 00:11:57,960 where enterprise players monopolize how AI can be 175 00:11:57,960 --> 00:12:00,800 deployed at scale. XAI is attempting to do the 176 00:12:00,800 --> 00:12:04,120 opposite by removing the human middleware and allowing Grok to 177 00:12:04,120 --> 00:12:07,800 act as the architect, the engineer, the tester, and the 178 00:12:07,800 --> 00:12:12,240 user interface designer. Teams within XAI are using 179 00:12:12,240 --> 00:12:15,640 custom prompts in reinforcement learning techniques, the train 180 00:12:15,640 --> 00:12:19,520 agents that iterate on their own mistakes and optimize outputs 181 00:12:19,840 --> 00:12:24,680 based on user like goals. So instead of simply being a 182 00:12:24,680 --> 00:12:29,040 coding assistant like Copilot, Grok is being trained to become 183 00:12:29,040 --> 00:12:32,480 a full stack developer and a product designer in one. 184 00:12:32,800 --> 00:12:37,200 It's kind of what Lovable does. Lovable dot dev, if you haven't 185 00:12:37,200 --> 00:12:41,320 checked it out, if you want to make a web app or anything in 186 00:12:41,320 --> 00:12:45,800 React, basically the time is now. 187 00:12:46,440 --> 00:12:48,840 Go to Lovable dot dev. I'm like, we're not sponsored by 188 00:12:48,840 --> 00:12:51,880 them, but check it out. I use them all the time. 189 00:12:52,040 --> 00:12:58,240 If I have AFO project, I give it the specs and of course I 190 00:12:58,240 --> 00:13:01,000 disclose that I'll be using AI to build some of the software 191 00:13:02,120 --> 00:13:06,920 and come up with with some great ideas Lovable has. 192 00:13:07,520 --> 00:13:11,320 And how funny is it though? Something like Lovable has its 193 00:13:11,440 --> 00:13:16,360 own ideas from the systems that it's already been implementing. 194 00:13:16,880 --> 00:13:18,520 It's learning as it goes as well. 195 00:13:19,360 --> 00:13:25,800 And XAI is, I wouldn't say catching up to it, not yet. 196 00:13:25,800 --> 00:13:28,240 If they could build this OS, if they can build a kind of a 197 00:13:28,240 --> 00:13:33,520 presentation layer of a Windows OS, Windows Cell OS, yeah, it'd 198 00:13:33,520 --> 00:13:36,800 be incredible. So the larger context though, of 199 00:13:36,800 --> 00:13:40,480 macro Hard includes a growing concern within Musk's network. 200 00:13:40,600 --> 00:13:44,400 The Microsoft's influence over AI infrastructure has become too 201 00:13:44,400 --> 00:13:48,840 concentrated, and by rebuilding key enterprise functions through 202 00:13:48,840 --> 00:13:52,840 autonomous agents at XAI, they want to demonstrate that open 203 00:13:52,840 --> 00:13:56,040 alternatives are possible. Could we, could they build this 204 00:13:56,600 --> 00:14:01,840 and make it sort of like Linux, an open sourced alternative to 205 00:14:01,840 --> 00:14:04,240 Windows that could run Windows games? 206 00:14:04,880 --> 00:14:08,520 I think that's where the logic goes from here. 207 00:14:09,240 --> 00:14:13,640 Elon, he's a big gamer. And I bet you guarantee you one 208 00:14:13,640 --> 00:14:17,600 of the one of the prompts from Elon to his engineering team was 209 00:14:18,280 --> 00:14:21,760 make it able to play games eventually. 210 00:14:22,040 --> 00:14:24,880 Like that's what we want. We want it to be a Steam killer, 211 00:14:25,120 --> 00:14:27,080 the Steam Deck. We want it to be a Windows 212 00:14:27,080 --> 00:14:30,680 killer and we want it to be able to play Windows games. 213 00:14:31,280 --> 00:14:32,320 Can they do that? I don't know. 214 00:14:32,320 --> 00:14:34,240 That's going to be crazy if they can. 215 00:14:36,480 --> 00:14:39,000 But Musk hasn't clarified yet if they're actually going to 216 00:14:39,000 --> 00:14:42,360 release this. But he suggested internally that 217 00:14:42,360 --> 00:14:47,760 such AI created software should run on a future Grok OS that 218 00:14:47,760 --> 00:14:49,800 avoids reliance on legacy platforms. 219 00:14:50,320 --> 00:14:53,040 Make everything open source, everything that he wanted to do 220 00:14:53,040 --> 00:14:56,000 with open AI. Think he gave him 50 million or 221 00:14:56,000 --> 00:14:58,120 something, 500 million. I can't remember the the price 222 00:14:58,120 --> 00:15:01,680 that he gave them or how much money he gave them, but I 223 00:15:01,680 --> 00:15:06,120 believe he's kind of hurt because they went from being 224 00:15:06,200 --> 00:15:12,200 open AI, open AI where they're going to share everything to Sam 225 00:15:12,200 --> 00:15:16,520 Altman getting ridiculously rich from this and all this 226 00:15:17,160 --> 00:15:21,080 stakeholders in this hoping that it everybody becomes ridiculous 227 00:15:21,080 --> 00:15:23,920 rich, ridiculously rich and the bubble doesn't burst. 228 00:15:25,400 --> 00:15:30,280 But as we see now, legacy systems like Google, Microsoft, 229 00:15:30,600 --> 00:15:35,600 all the initial Internet bubble Internet companies that are 230 00:15:35,600 --> 00:15:38,480 still around, you know that Yahoo is still around. 231 00:15:39,680 --> 00:15:43,160 Yeah, it's hard to see something like that unless you're into 232 00:15:43,160 --> 00:15:48,880 sports and like sports betting and stuff like that or fantasy 233 00:15:49,280 --> 00:15:52,080 sports. But yeah, Yahoo is still around. 234 00:15:52,640 --> 00:15:54,920 It's a thing. And you know what? 235 00:15:55,000 --> 00:15:57,600 It's a legacy company. Same with Microsoft, same with 236 00:15:57,600 --> 00:15:59,080 Google. They're legacy companies now. 237 00:15:59,440 --> 00:16:01,080 They do think different, do things differently. 238 00:16:01,080 --> 00:16:03,160 They don't think of the open source anymore, really. 239 00:16:03,560 --> 00:16:07,160 They have a small arm of open source, but if Elon can make a 240 00:16:07,160 --> 00:16:10,360 whole operating system and an office software suite open 241 00:16:10,360 --> 00:16:13,560 source, that would be ridiculous. 242 00:16:13,640 --> 00:16:17,960 If they release this, they'll be nuts, but I don't know if I'd 243 00:16:17,960 --> 00:16:20,000 trust it. Let me know. 244 00:16:20,000 --> 00:16:22,600 Again, like I said before, I need to know what you think. 245 00:16:23,480 --> 00:16:27,600 Let me know what you think about this because it sounds a little 246 00:16:27,600 --> 00:16:33,400 too personal. AI building the software, 247 00:16:33,840 --> 00:16:37,960 reiterating the software, revising the software, make an 248 00:16:37,960 --> 00:16:41,160 AI infrastructure so it keeps building and rebuilding 249 00:16:41,160 --> 00:16:42,320 software. I mean, sure, it's going to be 250 00:16:42,320 --> 00:16:48,680 great software eventually, but do you want an AI in charge or 251 00:16:48,680 --> 00:16:50,760 you don't want people in charge? I it's up to you. 252 00:16:50,800 --> 00:16:55,080 I mean, I have no problem with AI building software. 253 00:16:55,720 --> 00:16:57,320 I use AI to build software all the time. 254 00:16:57,800 --> 00:17:00,480 I think it's great. It's a great cheat code. 255 00:17:00,600 --> 00:17:04,240 You unlock certain things within the software development 256 00:17:04,240 --> 00:17:09,319 platform, like Lovable, and it just goes wild. 257 00:17:11,000 --> 00:17:15,440 You can create some really incredible things, Very complex 258 00:17:15,440 --> 00:17:19,079 things too. Very, very complex things. 259 00:17:19,079 --> 00:17:24,920 I've built some full blown apps, release them to the public, and 260 00:17:24,920 --> 00:17:28,200 they're crazy. You know? 261 00:17:28,720 --> 00:17:32,120 XAI or Elon has not commented publicly on the details of this, 262 00:17:32,120 --> 00:17:38,560 though you know there are. The internal prototypes already 263 00:17:38,560 --> 00:17:41,840 performed basic functions that mirror those of Office and 264 00:17:41,840 --> 00:17:44,440 Teams. One prototype allows Grok to 265 00:17:44,440 --> 00:17:49,360 compose documents real time and revise them based on simulated 266 00:17:49,360 --> 00:17:54,800 team collaboration using a chat interface powered by agent 267 00:17:54,800 --> 00:17:59,160 driven back ends. This is also going to serve a 268 00:17:59,440 --> 00:18:01,640 deeper technical purpose for XAI. 269 00:18:02,120 --> 00:18:05,520 By forcing its models to handle complex software tasks without 270 00:18:05,520 --> 00:18:09,200 pre-existing APIs or templates, the company is generating a 271 00:18:09,240 --> 00:18:13,800 stress testing environment for the new versions of Grok. 272 00:18:13,960 --> 00:18:16,640 The new models. These future systems will likely 273 00:18:16,640 --> 00:18:20,920 integrate self repairing code. Oh my, this is this is getting 274 00:18:20,920 --> 00:18:23,640 scary. Autonomous workflow planning and 275 00:18:23,640 --> 00:18:27,080 adaptive UI generation features that Musk believes are necessary 276 00:18:27,600 --> 00:18:34,640 for AGI to become practical. It has to think for itself self 277 00:18:34,640 --> 00:18:37,760 repairing its code. It's kind of like, how do you 278 00:18:37,760 --> 00:18:42,480 think you you make mistake and because of that mistake, you fix 279 00:18:42,480 --> 00:18:44,680 it, right? You fix whatever mistake it was, 280 00:18:45,120 --> 00:18:46,400 you make yourself a better person. 281 00:18:47,480 --> 00:18:50,840 Self repairing code, we do that, we already do that. 282 00:18:51,200 --> 00:18:59,400 We repair the wrongdoings. So once this autonomy comes to 283 00:18:59,400 --> 00:19:02,840 fruition, AGI will become more practical. 284 00:19:05,960 --> 00:19:09,960 Elon is aiming for broader things. 285 00:19:11,720 --> 00:19:15,560 He's going to be figuring out how software will be written in 286 00:19:15,560 --> 00:19:20,880 the future, desktop software or software that creates desktops 287 00:19:20,920 --> 00:19:24,920 and creates operating systems. And if XAI can figure this out, 288 00:19:26,320 --> 00:19:32,480 of course Tesla bots, Optimus, Tesla vehicles, etcetera, will 289 00:19:32,480 --> 00:19:38,320 all use this in the back end to make their code faster, more 290 00:19:38,320 --> 00:19:43,280 reliable, stronger, more productive. 291 00:19:43,680 --> 00:19:49,240 And guess what bye bye all software engineers, everyone's 292 00:19:49,240 --> 00:19:55,520 getting fired. So future AI with croc no 293 00:19:55,520 --> 00:19:59,320 people, no no no engineers anymore. 294 00:19:59,320 --> 00:20:02,000 There will be some high level engineers but all low level 295 00:20:02,000 --> 00:20:06,640 after the high like anywhere from medium till in turn are 296 00:20:06,640 --> 00:20:09,800 gone. You know, but I think for the 297 00:20:09,800 --> 00:20:15,560 time being, expect to not have your $165,000 starting salary. 298 00:20:15,560 --> 00:20:20,080 If you are a MIT grad. Expect to work at Chipotle. 299 00:20:20,600 --> 00:20:26,600 Because once things like this hit the market, that career is 300 00:20:26,600 --> 00:20:29,920 dunzo. And once Grok can figure out how 301 00:20:29,920 --> 00:20:36,080 to take out the C-Suite of people, and once other AIS can 302 00:20:36,080 --> 00:20:37,880 figure out how to take out the C-Suite and build their own 303 00:20:37,880 --> 00:20:41,040 software to to destroy the C-Suite, it's all over. 304 00:20:42,440 --> 00:20:48,280 The business will be handled by mostly AI, and then we'll just 305 00:20:48,480 --> 00:20:51,400 buy the things from the AI. It's going to be wild. 306 00:20:54,240 --> 00:20:56,000 Hey, thank you so much for listening today. 307 00:20:56,000 --> 00:20:58,000 I really do appreciate your support. 308 00:20:58,000 --> 00:21:00,840 If you could take a second and hit this subscribe or the follow 309 00:21:00,840 --> 00:21:03,040 button on whatever podcast platform that you're listening 310 00:21:03,040 --> 00:21:05,680 on right now, I greatly appreciate it. 311 00:21:05,680 --> 00:21:08,360 It helps out the show tremendously and you'll never 312 00:21:08,360 --> 00:21:11,320 miss an episode. And each episode is about 10 313 00:21:11,320 --> 00:21:13,920 minutes or less to get you caught up quickly. 314 00:21:14,320 --> 00:21:17,080 And please, if you want to support the show even more, go 315 00:21:17,080 --> 00:21:22,360 to patreon.com/stage Zero. And please take care of 316 00:21:22,360 --> 00:21:24,680 yourselves and each other. And I'll see you tomorrow.